Participant Enrollment

Patients with PAH presenting for routine follow-up at the Pulmonary Hypertension Comprehensive Care Center at Columbia University Irving Medical Center between September 1, 2016, and July 31, 2017, were recruited during routine visits. Inclusion criteria included patients with group 1 PAH, defined as having a mean pulmonary artery pressure of greater than 25 mm Hg with pulmonary artery wedge pressure less than 15 mm Hg by cardiac catheterization and hemodynamic studies. Participants were aged at least 13 years at the time of enrollment and spoke English. Patients' World Health Organization (WHO) functional class (FC) was required to be I to III at the time of enrollment. Participants were excluded if they had a history of non-compliance with follow-up visits, reported use of any wearable physical activity tracker in the last 6 months, had WHO FC IV disease, or were unable to provide informed consent/assent. Participants provided informed consent or assent (based on age) and completed a Health Insurance Portability and Accountability Act of 1996 (HIPAA) form as required by the Columbia University Human Research Protection Office. Before the initial visit, each patient was given a Fitbit Charge HR activity monitor (Fitbit, Inc) with instructions on usage and synchronization to the patient's smartphone device or computer. The Fitbit Charge HR is a wireless, wrist-worn, triaxial accelerometer. A proprietary algorithm translates raw acceleration signals into steps and activity levels. It estimates steps, heart rate, activity level, and energy expenditure each minute.¹⁷

The Fitbit device was chosen given its ease of use, easy availability, and greater accuracy than that of other commercially available activity monitors.¹⁸ Patients were given instructions on how to upload data to a cloud interface. These data were downloaded by research personnel to a deidentified spreadsheet. The study protocol was approved by the institutional review board at Columbia University (protocol No. AAAQ9188).

Patient Characteristics and Functional Capacity Variables

Date of birth, sex, ethnicity, race, weight in kilograms, body mass index (BMI), WHO FC, and WHO diagnostic classification were charted for each patient. At the initial visit, all patients underwent a CPET and 6MWT. Peak oxygen consumption (VO₂ max) (in mL/kg/min), percentage predicted VO₂, and meters walked on 6MWT were used in the final analysis.

Steps per day and sedentary time per day (in minutes) were obtained from each Fitbit device, beginning from the day after the initial visit. Sedentary activity, which includes activities such as sitting quietly, sleeping, and watching television, is defined by Fitbit as 1.0 to 1.5 metabolic equivalents. These 2 variables were used, given their relative accuracy compared with other Fitbit data, such as distance walked per day, which may vary by stride, or calories consumed per day, which may vary by patient weight. Only days on which more than 500 steps were recorded were included for statistical analysis, as descriptive analysis of steps per day suggested that days with fewer than 500 steps might represent substantial outliers from technical factors of the device (eg, insufficient detection of steps, errors in uploading, or failure to wear the device) that did not truly reflect the patient's disease burden. For each patient, 7 consecutive compliant days of activity were used in the final analysis to maximize the likelihood of consistent wear time.

Transthoracic Echocardiography Variables

At the initial visit, patients underwent a clinically indicated transthoracic echocardiogram (Phillips iE33, Philips Healthcare). Qualitative global RV function (normal or mild, moderate, or severe dysfunction) was obtained from the echocardiogram report. If technical limitations or image resolution precluded any of the echocardiographic measurements being made, that participant was excluded from the study.

Tricuspid regurgitant peak velocity (in m/s), tricuspid annular plane systolic excursion (TAPSE; in cm), and tricuspid annular peak systolic velocity (TAPSV) by tissue Doppler (in cm/s; Fig. 1A); RV myocardial performance index by tissue Doppler (myocardial performance index [MPI], Fig. 1A); and pulmonary artery acceleration time (PAAT; in ms; Fig. 1B) were measured by one of the authors (J.L.W.), who was blinded to previous measurements and clinical data. These previously validated measurements are abnormal in patients with PAH as a result of abnormal RV myocardial dynamics.¹⁹ In addition, PAAT uses the flow velocity characteristics through the RV outflow tract for a quantifiable measure of pulmonary circulatory mechanics, and PAAT has been shown to be inversely proportional to both mean pulmonary artery pressure and indexed pulmonary vascular resistance on cardiac catheterization for both adults and children.^20,21 All parameters were measured in triplicate and averaged.

View Figure

• Download as Powerpoint
• Download Figure

Right ventricular global longitudinal strain (GLS) and strain rate (SR) were measured retrospectively using TomTec multivendor software version 4.6 (TomTec Imaging Systems). A standard 2-dimensional image of the apical 4-chamber view was used to assess RV longitudinal strain. The quality of the tracking was ensured for each segment and adjusted as necessary. The peak longitudinal strain and SR were recorded. All measurements were performed by 1 author (J.L.W.) on 3 separate occasions separated by several weeks, and this author was blinded to previous measurements and clinical data. Ten studies were chosen at random, and measurements were performed by a second reader (N.P.) to determine interobserver reliability.

Statistical Analysis

Data are presented as mean (SD) or median (interquartile range [IQR]) for continuous variables and frequency counts (percentage) for categorical variables. Comparisons of echocardiographic data with WHO FC groups were conducted using analysis of variance or Kruskal–Wallis tests, depending on the distribution of the data. Pearson (r) and Spearman (r) correlations were calculated for continuous variables. Intraclass correlation was used to assess intraobserver reliability for MPI and PAAT and both intra- and interobserver reliability for RV GLS/SR. An intraclass correlation threshold of 0.75 was used to indicate good intraobserver and interobserver reliability.²² If the intraclass correlation exceeded 0.75, the primary reader's (J.L.W.) measurements were used for statistical analysis, including those instances in which the 2 readers disagreed. Significant P values were defined as < .05. Statistical analyses were performed using Stata14 (StataCorp).

Patient Characteristics

Thirty-two patients were initially enrolled in the study, 22 patients of which had complete data for analysis. One participant withdrew after enrollment because of developing a rash, and 3 patients did not have 7 consecutive compliant days of Fitbit data. One participant did not have an exercise study at the time of enrollment. One participant did not have images suitable for MPI or PAAT measurements, and 4 participants did not have images that were suitable for strain analysis.

There were equal numbers of males and females in the cohort (Table I). The median (IQR) age of the cohort was 27 (15–46) years, with 59% of patients older than 30 years. The mean (SD) BMI was 23.4 (6.2) kg/m². Nearly half of the patients were White (n = 9 [41%]), and half had WHO FC II symptoms (n = 11 [50%]). The mean (SD) number of steps per day for the entire cohort was 6,237 (3,498) steps. The median (IQR) sedentary time per day for the entire cohort was 962 (676–1,262) minutes.

TABLE I Patient Demographics and Clinical Characteristics

View Fullscreen

Echocardiography Data by WHO Functional Class

Average peak tricuspid regurgitant (TR) velocity increased by WHO FC and was nearly 50% higher in the WHO FC III than the WHO FC I group (2.9 [0.74] m/s in WHO FC I, 3.8 [0.52] m/s in WHO FC II, 4.5 [0.59] in WHO FC III; P < .01; Fig. 2A). Average RV GLS worsened by WHO FC (−20.6% [3.8%] in WHO FC I, −16.3% [5.2%] in WHO FC II, −14.1% [3.3%] in WHO FC III groups), although this did not reach statistical significance (P = .07; Fig. 2B). Similarly, average RV septal strain worsened with higher WHO FC (−23.8% [5.1%] in WHO FC I, −19.3% [5.6%] in WHO FC II, −16.5% [6.8%] in WHO FC III groups), although this did not reach statistical significance (P = .07). In considering a potential confounding effect of posterior bowing of the interventricular septum on the relationship between WHO FC and strain, although there was an association between increasing posterior bowing and worsening septal strain (P = .04), an association was not found with WHO FC (P = .86) or RV GLS (P = .09). TAPSE, TAPSV, RV MPI, PAAT, and RV free wall strain did not significantly change across WHO FC groups.

View Figure

• Download as Powerpoint
• Download Figure

Correlation Between Echocardiography Data and Functional Capacity

Tricuspid regurgitant peak velocity had a negative correlation with 6MWT performance (r = −0.58, P = .02; Fig. 3A), average steps per day (r = −0.59, P = .03; Fig. 3B), and VO₂ max on exercise testing (r = −0.55, P = .03). Conversely, TR peak velocity had a positive correlation with median sedentary time on Fitbit (r = 0.64, P = .02; Fig. 3C). Pulmonary artery acceleration time positively correlated with VO₂ max on CPET (r = 0.63, P < .01; Fig. 3D). Right ventricular TAPSE, TAPSV, MPI, GLS, free wall strain, septal strain, and global SR were not significantly correlated with accelerometry or functional capacity testing (Table II).

TABLE II Correlation Among RV Echocardiographic Measurements, Accelerometry, and Functional Capacity Testing ^a

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Intraobserver and Interobserver Reliability

Intraclass correlation for intraobserver reliability was 0.82 for MPI (95% CI, 0.63–0.92), 0.91 for PAAT (95% CI, 0.76–0.96), and 0.86 for RV GLS and SR (95% CI, 0.52–0.89). Intraclass correlation for interobserver reliability for RV GLS was 0.78 (95% CI, 0.51–0.92).